The Nineteenth Byte

quartata

03:00

C++ lambdas fit just fine in a std::function

Function pointers are a thing for C statement block functions

Qwerp-Derp

Does the syntax above work

quartata

I think you grossly underestimate how powerful modern C++ is. That's why we're not compiling to LLVM.

Qwerp-Derp

A(m, n) -> [m = 0: n + 1
           [n = 0: A(m - 1, 1)
           [A(m - 1, A(m, n - 1))

Downgoat

Ok well if you're going to C++ that's different

quartata

Not really. Function pointers exist in C too.

Downgoat

03:02

@Qwerp-Derp I'm not a fan at the lack of matchinf braces and less whitespace honestly

Qwerp-Derp

Golfed:

quartata

Most of the transpiled code will be like C. We'll probably need to end up plundering C++'s stdlib for implementation details (particuarly shared:) is all.

Qwerp-Derp

That's ungolfed

A(m,n)->[m=0:n+1,n=0:A(m-1,1),A(m-1,A(m,n-1))

Should I change to curly braces?

Downgoat

@quartata ok that would work

quartata

@Qwerp-Derp Yeah. Looks more like a piecewise function like that

Downgoat

03:05

Still not sure on how you will manage to implement polymorphism in C but OK

Qwerp-Derp

piecewse?

What

Oh

quartata

spelling is hard

H Walters

@Downgoat it's not really all that hard... you just follow a slightly different idiom

Qwerp-Derp

A(m, n) -> {m = 0: n + 1
           {n = 0: A(m - 1, 1)
           {A(m - 1, A(m, n - 1))

quartata

@Downgoat Polymorphism was a bad word. Envision it more like a struct with a union and an enum if you want

Downgoat

03:07

@quartata will primitives be wrapped in PytekObjects then?

Qwerp-Derp

So like that

It looks nice

quartata

@Downgoat Yes.

H Walters

@Downgoat "virtual functions" map through function pointers. "Method calls" are just functions, where you explicitly pass the equivalent of "this" pointers.

Downgoat

@HWalters well you'd have to go and kinda manually reimplemment scopes in a sense.

H Walters

Setting your "virtual function table" is a matter of picking the right C function to allocate/initialize your object

Downgoat

03:08

@quartata ok cool, thanks

H Walters

Well, yeah, you do, but it's actually not that rare of a thing, and isn't terribly difficult... you just follow a C-like idiom

Qwerp-Derp

Does the syntax look alright?

Wait so it's called a piecewise function?

Downgoat

Difficulty isn't a problem, but any naive solution to such would be incredibly slow to my understanding

@Qwerp-Derp the "if" structure? Yeah

Qwerp-Derp

@Downgoat Huh

The more you know

H Walters

@Downgoat The straightforward way to implement this type of thing should be pretty much equivalent to virtual functions in C++ in terms of performance

Qwerp-Derp

03:12

Also what does Haskell's pattern matching have to do with this

Downgoat

@Qwerp-Derp it's basically exactly what you're doing

Qwerp-Derp

So is Haskell's pattern matching essentially the same as my syntax for CMQ?

CMQ is my lang btw

Downgoat

Wrapping primitives seems to be slow but I guess any decent compiler should be able to do basic optimization to speed up

@Qwerp-Derp no but example is:

sign x |  x >  0        =   1
       |  x == 0        =   0
       |  x <  0        =  -1

Pavel

@Qwerp-Derp you still haven't actually shown any kind of sequence in CMQ.

Qwerp-Derp

What do you mean @Pavel

Pavel

03:19

You said it's based on sequences but you haven't shown us any behavior like that.

Also, come on guys, we need more than four languages in the time capsule.

quartata

@Downgoat Accessing class members is not slow

Downgoat

@quartata huh?

Pavel

@quartata go add your language.

docs.google.com/forms/d/…

quartata

You said wrapping primitives is slow. Accessing class members isn't slow

@Pavel isn't my language, isn't my place to say

Not confident we can make 2018 anyways

ais523

accessing class members isn't slow, unless they're virtual

virtual accesses are fairly slow compared to static ones

quartata

03:27

At least not a full release. Our STDLIB is going to be a nuisance

H Walters

I've got an idea for a useful language, but the idea isn't fully developed

Downgoat

@quartata I'm not talking about that rather operator overloading and all, which I assume you'll have. Will all the runtime checks due to dynamic types, that would pose a bit of a problem for speed, no?

H Walters

Idea's based on a typical workflow... oft times when analyzing data I'll combine vim with binutils and Excel...

quartata

@Downgoat Dynamic type checking is only necessary if the type isn't specified. If you care about speed you'll do that. If you don't it will still be better than interpreted

H Walters

...the idea would be to replace the latter two with an analysis language based on columns/array data types

Downgoat

03:32

@quartata still be better than interpreted? Hardly when regarding dynamic types, as JIT runtime optimizations will kick in

quartata

@Downgoat Interpreted still has to parse and Cython isn't perfect.

Cython won't match C even though it will be orders of magnitude faster than pure Python

Pavel

Well, you have a year to make it @HWalters

Eᴀsᴛᴇʀʟʏ Iʀᴋ

VTD this answer? it's blatant NAA

quartata

@HWalters Have you played with APL or J or K before? Very array based.

Pavel

I have not heard those acronyms before.

H Walters

03:41

@quartata: No, but I'm aware of that

quartata

@Pavel They're languages.

H Walters

Yeah, I know that, and that they are array based

quartata

not you :P

H Walters

Oh, sorry

The not-quite fleshed out language I have in mind will work like "relations"; similar to columns in a spreadsheet (or tables in a database)

Downgoat

@quartata I think you are overestimating how much time parsing takes by a collosal amount. 0.5ms more on boot-up time is worth having a significant speed boost at run time

Wheat Wizard

03:49

Some one should make a meta-cops-and-robbers where cops post cops-and-robbers problems and robbers crack them by finding cryptographically secure ways to make answers.

Conor O'Brien

@Pavel can someone submit multiple things?

quartata

@Downgoat I think you're overestimating what the speed boost is or that it would even exist at all

First off we're not doing duck typing.

That's pretty much it actually. The only optimization I can thinl of granted by JIT is making sure the class tree isn't cold.

It's one lookup either way

Small price to pay if you're not specifying types.

Now if we were doing duck typing that would be a different story -- checks would have a lot of hot spots that could be optimized JIT

Destructible Watermelon

hi

quartata

@WheatWizard That kind of sounds like rules lawyering and/or just rolling your own crypto which I'm not sure how I feel about

Wheat Wizard

@quartata What do you mean by "rules lawyering"?

quartata

04:00

i.e finding cheap cryptographic methods not excluded in the problem description

Wheat Wizard

Isn't that always a problem with cops-and-robbers?

ais523

it depends on what the problem is

if it's "figure out the language this program is written in" then crypto won't help much

Wheat Wizard

I didn't intend the challenge as an actual question. I just found it funny how many cops-and-robbers problems can be solved in cryptographically secure methods.

Conor O'Brien

enter this into this when it works

Pavel

04:22

@ConorO'Brien sure, the challenge will rely on there being many but there's not a lot there now.

I made it so anyone can see the results to the time capsule form btw. The description field is optional.

Destructible Watermelon

seriously what happened to calesyta?

Pavel

04:39

What's that?

Conor O'Brien

@Pavel I don't seem able to see the results.

Sherlock9

@Mego Thanks

Pavel

I'm not really sure why...

Conor O'Brien

do you have the form connected to a google spreadsheet?

Pavel

Try this link docs.google.com/forms/d/…

Conor O'Brien

04:43

just brings me to the form

Pavel

And no, I can't figure out how on mobile.

@Downgoat are you responsible for the downsheep submission, or is that an actual user?

Downgoat

Did SE fix MathJax? CR is running it so it seems likely that it is now working

Pavel

Ok, have spread sheet: docs.google.com/spreadsheets/d/…

@ConorO'Brien ^

I haven't come up with a name for my language yet.

Downgoat

@Pavel Perhaps make it public so we can all see?

Pavel

Dur...

Oops.

drive.google.com/…

Spreadsheet works now, in theory. @Downgoat @ConorO'Brien

Conor O'Brien

05:04

alright looks good

New Sandboxed Posts

0

A: Sandbox for Proposed Challenges

Compute the Wilson score interval The Wilson score interval is a confidence interval of the probability of success, based on the proportion of successes in a set of Bernoulli trials (a Bernoulli trial is a trial in which exactly two independent outcomes are possible). The interval is given by th...

Anonymous

@NewSandboxedPosts Damn that was fast

wat

what do you think of my anguage idea

Pavel

What is it

wat

look at t he spreadsheet

betseg

05:16

@Mego post says 6 min ago, bot messaged at 08:09

Wtf

Pavel

That's... Uh... Interesting. Are there eight Whitespace characters in ascii?

Anyway, that'll take a good fifteen minutes to implement. You can make more than one submission btw.

ais523

@Pavel there are a few that are controversial: in order of least controversial to most, it has space, linefeed, tab, formfeed, carriage return, vertical tab, a few separators like FS and GS, and NUL

interestingly, there's an actual "newline" character in Latin-1 (and thus Unicode) that's distinct from any of the ASCII-7 whitespace

(and of course there's a bunch more possible whitespace in Unicode)

I think by most people's counts, though, the number of whitespace characters in ASCII is in the 4 to 6 range

Pavel

Well, wat is gonna need eight to make a bf derivative.

At least if I understood right that is basically alphafuck but white space.

wat

05:33

@Pavel actually base 3

I only use space tab and enter

Downgoat

@noɥʇʎԀʎzɐɹƆ do you know of any officialish tool/standard to generate python code (i.e. AST -> Python)? Language doesn't really matter

Pythons ast module doesn't seem to let me make an AST but I might be wrong

I am also not sure if that sypoorts like Cython

wat

basically it will be two trits for one of 8 characters

and a nop

Downgoat

Like JS has an extendable AST format w/ code gen such as babel

Anonymous

@Downgoat ast.parse?

Anonymous

>>> ast.dump(ast.parse('print("Hello, World!")'))
"Module(body=[Expr(value=Call(func=Name(id='print', ctx=Load()), args=[Str(s='Hello, World!')], keywords=[], starargs=None, kwargs=None))])"

Pavel

05:47

Not sure if I'm happy more people are participating in the time capsule, or disappointed there will soon be yet another trivial bf derivative out there.

Destructible Watermelon

I'm working on my koth

H Walters

Oh that's a great idea... a bf derivative that just shuffles the characters around

For the sole purpose of making it really hard to polyglot it with bf

wat

What should I use as a nop in Brainfuck that isn't whitespace?

Pavel

Will everything else be a syntax error?

Destructible Watermelon

@wat G

don't ask why

Pavel

05:49

My vote is for carriage return.

wat

@Pavel @DestructibleWatermelon everything but white space is ignored in source, but I need something to use as nop in the intermediate brainfuck code.

Downgoat

@Mego That uses existing python code however

Anonymous

> A friendly message from The Name of All that is Good and Holy: Hello! This is The Name of All that is Good and Holy. Before you think of making your own extension to this language, here's another idea for you to consider: Don't! For the love of me, please don't make another derivative of this language! Unless you have something truly new to contribute, it's been done a million times before and will probably make it worse than the original.

2

Destructible Watermelon

I still say you should use G as nop

Pavel

The polyglot argument actually works here.

betseg

05:50

@wat |^=\%_&#@*"';!?/ One of these

wat

Maybe I should use the character "="

ugh you beat me to it

Destructible Watermelon

@wat s/=/G/

Anonymous

@Downgoat True. What you're thinking of (generating Python code from an AST you made yourself) is probably not reasonable.

wat

@DestructibleWatermelon whatever

betseg

@DestructibleWatermelon random≠funny

Pavel

05:51

That and I don't know how to remove submissions from Google forms.

wat

@Pavel delete the row in the spreadsheet

Destructible Watermelon

hmmm, I'm not going to need this class am I?...

Pavel

Won't it just be regenerated, being linked to the form?

Downgoat

@Mego Ok, I guess I could implement into my python fork

muddyfish

@Mego honestly it's not too bad as long as it's relatively simple

Pavel

05:53

If you actually want me to remove it I'll do it when I'm at a computer.

Anonymous

This might be useful

H Walters

Yeah, when you have access to the internet

Destructible Watermelon

I have this as a class and just realised I'm probably not going to need to extend it anyway

class Artifact:
    def __init__(self, x, y):
        self.x, self.y = x, y

time for a list I guess

Pavel

@DestructibleWatermelon are you working on a language?

Destructible Watermelon

no a KOTH

Pavel

05:56

Oh right.

Destructible Watermelon

any function I could think of isn't really relevant

and could just be outside

Pavel

The speed sheet crashes in seconds btw. It took me a good several minutes to make it public.

wat

@Pavel no

H Walters

What a fun typo to read: "The speed sheet crashes in seconds"

Or was that intended?

Pavel

Hah, didn't notice.

Downgoat

06:01

@Mego O_o what's the use of AST parser without code gen...

Anonymous

@Downgoat What do you mean?

betseg

06:23

Packaging of my soap says it's not edible

Destructible Watermelon

@betseg did you realise too late?

betseg

I wonder who tried to eat and sued Dove

Destructible Watermelon

probably not dove specifically

Pavel

At this point companies probably put that label on everything just in case.

betseg

Tru dat

Pavel

06:32

I still need a name for my language... The name is always the hard part for me.

I don't know how you guys do it.

Dennis

With one exception, I just pick something edible.

Wait, two exceptions.

H Walters

Make up a symbol and call it a name

Pavel

Tuna sandwich it is.

Actually, I'll call it 🍣, pronounced sushi.

Even though there probably won't be emoji in the syntax.

H Walters

Sushi can be spelled with Martian letters

Pavel

How does that work?

🍣 because it's inspired by ><> btw.

Qwerp-Derp

06:46

I think I'll post a snippet for the Processing thing every day

H Walters

Excert: "From India to the Planet Mars" See figure 24.

Pavel

Uhh... Is the link to the form on the right side bar? If it is, can someone paste it here for me cause I'm on mobile.

Qwerp-Derp

@Dennis Wait what languages have you made

Pavel

Please? I can't access the form without signing out or I get taken to edit view.

Qwerp-Derp

I can't think of any interesting 1-char snippets for Processing

H Walters

06:49

Why 1-char snippet? Didn't that thread just get redefined as a wiki/showcase?

Qwerp-Derp

Oh

Destructible Watermelon

@HWalters ^^

Dennis

@Qwerp-Derp ShapeScript; Help, Wardoq!, Bubblegum, Jelly, and Sesos.

Pavel

iirc my limited processing knowledge: ' in many languages defines a string literal, bit in processing instead defines a char literal.

Is sesos a food I haven't heard of?

Qwerp-Derp

@Pavel It means "culinary brains" apparently

See Sesos's wiki

Pavel

06:51

Til

Destructible Watermelon

also looking at github there are only 2 food esolangs there

Qwerp-Derp

@Pavel Isn't that the same in Java though?

Dennis

@DestructibleWatermelon Bubblegum isn't on GitHub.

Destructible Watermelon

ah

Pavel

Well, few languages can have something entirely unique in one byte.

Destructible Watermelon

06:53

I still say that 3 versus 2 doesn't really count as a rule

Qwerp-Derp

actually processing tbh isn't different enough from Java to warrant a different post

Pavel

Dennis, you should enter the time capsule.

Destructible Watermelon

@Dennis did you say you were going to make a jelly 2 or something

because if you are you should call it beans

H Walters

No he can't

Pavel

Make it for the time capsule

H Walters

06:55

"Beans" is implicitly reserved for something java related

As is every word that's remotely coffee related

Pavel

Oh right

Qwerp-Derp

NetBeans

Destructible Watermelon

hmmm wait my koth has a mechanic I want to implement but it doesn't quite work with my current system

Qwerp-Derp

I kinda want to do a language that I can make one snippet for every day for a month

Destructible Watermelon

@Qwerp-Derp I'm working on the koth some more. I didn't know if you wanted to help maybe

Qwerp-Derp

06:57

@Dennis Does the language have to be implemented, or can it be just an idea for what's to come?

@DestructibleWatermelon Uhhh sure what do you need help with?

Destructible Watermelon

@Qwerp-Derp just general stuff

Qwerp-Derp

Like what?

Destructible Watermelon

I have the ideas down now I think

I just need to implement it mostly

Qwerp-Derp

Chuck us list of ideas

I will try to think of loopholes so that you can patch it up

Destructible Watermelon

care to join the room?

Qwerp-Derp

06:58

Sure

Hmmm... I'll add Doxical in

Pavel

Or a Github link

You'll need one at some point for the controller

Qwerp-Derp

@Dennis?

Dennis

@DestructibleWatermelon At some point. TIO comes first.

Pavel

You could just do CMQ for the showcase.

Dennis

@Qwerp-Derp It should be implemented.

Pavel

07:00

Aww.

Destructible Watermelon

@Qwerp-Derp are you joining the room?

Pavel

Seriously, does anyone have the link to the form?

Qwerp-Derp

@Pavel That's what I want to do

H Walters

@Pavel This one?

Qwerp-Derp

@Dennis Should or must?

Pavel

07:02

Thank you

I only have an edit link

Dennis

@Qwerp-Derp We define languages by their implementation. If there's no implementation, as far as PPCG is concerned, there is no language.

Qwerp-Derp

Orrite

Then let's get crackin' to make CMQ!

zyabin101

@Qwerp-Derp Must.

Pavel

I'm still waiting for a useful reply to this meta post.

wat

working on WSF right now

zyabin101

07:14

Whoa now, Stack Exchange required me to vote on questions. :)

Just because I haven't participated in a while XD

wat

@zyabin101 they don't "require" you

they just pop up a message

zyabin101

Yup.

Qwerp-Derp

Is anyone interested in helping to make docs for CMQ?

zyabin101

CMQ?

Pavel

His language

Qwerp-Derp

07:25

Charles Montgomery Quince, my language

zyabin101

:D

Pavel

C. Quince

zyabin101

"Hello, World!"?

Pavel

Sequence

Qwerp-Derp

@zyabin101 What

zyabin101

07:25

@Pavel You get it. :)

Destructible Watermelon

is the acronym CMQ canonical?

Qwerp-Derp

@DestructibleWatermelon That's probs going to be the extension of the file

zyabin101

@Qwerp-Derp Which program makes Charles output Hello, World!?

Qwerp-Derp

So like hello.cmq

It's based on sequences, I have to figure out ASCII support

So is anyone interested?

Pavel

Interpreted, transpiled, or compiled?

H Walters

07:27

I have experience using ASCII characters

Qwerp-Derp

@Pavel It's probs going to be interpreted

I'm thinking of making it in PegJS

Because it's easy to make parser stuffs in

Pavel

Don't know JS, not gonna learn now.

So probably can't help you there

zyabin101

@Qwerp-Derp _o/

Running "c:\fpc\3.0.0\bin\i386-win32\cradle.exe "
HelloWorld
Running "c:\fpc\3.0.0\bin\i386-win32\cradle.exe "
123

LMAC's cradle works.

Qwerp-Derp

@zyabin101 Yey

Or I could write it in Python or something but I need a proper parser and stuff

Pavel

Most interpreters these days seen to be written in python.

Qwerp-Derp

07:35

Should I use pegjs or pegcoffee?

zyabin101

@Pavel For recreational langs, yes.

Practical langs tend to use more advanced tools like C or Java.

H Walters

I've written a few in C++

zyabin101

@Qwerp-Derp Whatever the heart feels! :D

H Walters

Which I'm sure surprises nobody here

Qwerp-Derp

I feel like I'm not good enough in C++ to make a full-fledged language

I like to fall back to higher-level languages

Pavel

07:37

Dynamic typing is useful for data structures.

Qwerp-Derp

Wait what is dynamic typing

Pavel

Unlimited precision ints is nice too.

@Qwerp-Derp variables don't have type declaration and type can change.

Like in JS.

H Walters

@Qwerp-Derp at first approximation it's typing that's handled at run-time

Qwerp-Derp

Oh

Pavel

For interpreters, it means you can have a list of variables without worrying about what those variables are.

Qwerp-Derp

07:40

I like dynamic typing - that's probably why I don't use low-level languages like C++ & Java

Pavel

In Java for example,I could make a hashmap for variables, but I'd have to put in some effort to store integers and strings.

Which is why I'm writing 🍣 in python.

Qwerp-Derp

Does anyone want to help in making some docs?

For CMQ?

zyabin101

_o/

Qwerp-Derp

@zyabin101 Yey

Wait zyabin did you work on VSL?

What was the site that you guys used for cloud docs stuff

Pavel

I guess I can help a bit.

But I have other things to do m

zyabin101

07:43

@Qwerp-Derp No.

@Qwerp-Derp IDK, but I'll try setting up a GitBook.

Qwerp-Derp

We mainly need ideas for the syntax

H Walters

Rough high level idea... completely separate your syntax from your semantics

Qwerp-Derp

Oh sketchboard

@HWalters What?

Pavel

Mathematica-like syntax. Obviously.

H Walters

Basically make a form where you can define code structures, having nothing to do with anything in the language (including builtins like "loops")

zyabin101

07:45

Running "c:\fpc\3.0.0\bin\i386-win32\translator.exe "
1
        d0 = 1

I wrote a working translator \o/

H Walters

...then use that structure to lay out everything

I've done that for a language before, and it makes for really quick implementation

Pavel

Object oriented or not?

H Walters

(no "a semicolon can go here because we're in a for loop, but not here because ...")

Qwerp-Derp

Oh hackmd

zyabin101

I'll still make a GitBook.

Qwerp-Derp

07:47

OK.

zyabin101

(Instead of 68000 asm, I use Python in my translator.)

Qwerp-Derp

hackmd.io/OwTgzGIEwCZQtAMwEYBZH1QBiwU3siLgKzy64THIDGxUiiUQA===

There's the specs thing

If anyone's interested

Pavel

Jesus... They managed to make hackmd mobile even worse

Qwerp-Derp

Whoa what happened

Pavel

I can no longer type

H Walters

07:52

Meh, time to head on...

Pavel

Wait no I found a setting to override their system for input

It works now

zyabin101

Running "c:\fpc\3.0.0\bin\i386-win32\translator.exe "
3+5
        d0 = 3
        d1 = d0
        d0 = 5
        d0 = d1 + d0
Running "c:\fpc\3.0.0\bin\i386-win32\translator.exe "
7-8
        d0 = 7
        d1 = d0
        d0 = 8
        d0 = d1 - d0
Running "c:\fpc\3.0.0\bin\i386-win32\translator.exe "
8
        d0 = 8
        d1 = d0

Error: an addop was expected

Pavel

Uhh...i noticed a n=(0||1) in there.

Destructible Watermelon

what should I call the koth characters?

Pavel

In most languages that means n==true

Destructible Watermelon

07:55

soldiers?

Pavel

You might want a different syntax there

Qwerp-Derp

What why?

What's wrong with that?

Pavel

That seems like a case for elementOf rather than or.

And || is pretty universal or notation

Destructible Watermelon

also its using equals for comparison not assignment

is an issue he may be thinking of

Transcript for